We have been studying for some years the compound insulator at the 5 end of the chicken beta globin locus. We showed that in addition to its ability to prevent enhancer-promoter interactions, mediated by CTCF, it is also able to block the hetrochromatinization of a reporter gene. This property is independent of CTCF, but we have shown that it does depend on the binding of two other proteins, USF1/2 and BGP1. We have studied the role of USF1/2 extensively and shown that as part of its insulator function it recruits a wide variety of histone modifying enzymes, which serve to maintain nearby histones in an active state, and prevent other, repressive, histone modifications from being introduced. Our results show that USF1 interacts in vivo with the vertebrate Set1 complex, which methylates histone H3 at lysine 4, as well as PRMT1, which methylates histone H4 at arginine 3. These are present in two separate, multicomponent complexes which appear to be localized to the insulator element through specific binding of USF1/2. Recent studies have focused on the role of PRMT1. We have shown that methylation of H4R3 by this enzyme is necessary for histone acetylation and gene activation. Studies have been extended to the roles of USF1/2 and PRMT1 in regulation of beta-globin gene expression, where they are found play an important role. We have also investigated the properties of BGP1, which binds to separate sites in the insulator. We find that the mouse homolog of BGP1 in mouse ES cells plays an important role in DNA methylation;in its absence methylation levels at critical sites genome-wide are depressed. We have shown that this is quite specific for loss of BGP1 binding at the insulator. We designed experiments to test whether Vezf1 binding to sites near an APRT reporter could affect methylation patterns, and found that in the absence of Vezf1 sites the region's DNA became methylated extensively;binding of Vezf1 inhibited methylation. We have made use of an ES cell line in which Vezf1, the mouse BGP1, is deleted. In collaboration with Dr. H. Stuhlmann (Cornell Medical College) we showed that in the absence of Vezf1 the DNA de novo methyl transferase, Dnmt3b, is down regulated. Wild type phenotype can largely be restored by introducing a Vezf1 expression vector into these cells. We have shown that Vezf1 binds in vivo to a site in an intron of the Dnmt3b gene. We have now extended our investigation in an attempt to understand how Vezf1 suppresses Dnmt3b expression. In these studies, we are asking whether Vezf1 may interfere with transcription elongation in such a way as to alter the production of specific splice variants of a gene, as suggested by the experiments described above with mouse ES cells.